I am using Data Envelopment Analysis for my research. Most of the inputs and outputs are derived from the past literature. However I want to test some additional factors which may be considered as Input or Output factor. I read about the 3-level fractional factorial design as suggested by Hiroshi Morita and Necmi K Avkiran in their 2008 paper "Selecting Inputs and Outputs in Data Envelopment Analysis by designing statistical experiments".
1. Can somebody provide clarifications on how to use this approach?
2. Is there any other approach to choose inputs and outputs?
Kumar Arijit
1. The 3-level fractional factorial design is a statistical method used in data envelopment analysis (DEA) to select inputs and outputs. It involves selecting a set of inputs and outputs that are believed to be relevant to the problem. These factors are then varied at three levels (low, medium, high) to create a set of experimental conditions. DEA models are then run on each experimental condition to obtain efficiency scores. The results are analysed using statistical methods to identify the factors that have the greatest impact on the efficiency scores. This approach allows for the identification of important inputs and outputs, as well as their optimal levels, in a relatively efficient and effective manner.
2. There are other approaches to selecting inputs and outputs in DEA. These are expert judgment, conducting a literature review, analysing correlation and regression analysis, and using principal component analysis (PCA) and factor analysis. It is generally sound to use a combination of methods to ensure a comprehensive selection process.
- Badges
- Science topic
- Similar topics
- Econ
- Innovation Economics
- Innovation